mirror of
https://github.com/tangerineturtle1/NX2navi.git
synced 2026-11-19 23:49:56 -04:00
Wait for the whole boot prewarm and show its progress on a boot screen.
The game used to start after at most 3 s of prewarm, so after a translator, GPU or driver change it met the rest of its shaders in play and stalled on each one. It now waits for all of it: a warm start is ready in about half a second and shows nothing, while a cold one shows a "Preparing shaders" screen in the menu's colours and the game's fonts, with the shader and then the pipeline count and a bar. The shader and pipeline prewarm report their numbers every 100 ms. The present path takes a frame without a game image for the screen. The RomFS opens on the first host-side read, so the screen's fonts load before the guest's first file access. NX2NAVI_PREWARM_WAIT_MS now defaults to 0 (no cap). With no progress for 60 s the game starts anyway and the prewarm finishes in the background. Measured with an emptied cache: 2,988 shaders translated and 2,399 pipelines built in 9.4 s behind the screen, then 0 ms shader and pipeline waits in Kokiri Forest. Warm: the game starts 0.47 s after launch. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
@@ -2574,7 +2574,26 @@ static void warm_builder(void *arg) {
|
||||
|
||||
static void lb_start(const uint32_t pr_hist[16], const uint32_t fs_hist[4]);
|
||||
|
||||
// vk_pipeline_prewarm_progress's numbers: the phase, and the pipelines finished (built or failed) of those queued, copied
|
||||
// while prewarm_run waits for its builders.
|
||||
static int32_t pw_phase;
|
||||
static int64_t pw_pipe_done, pw_pipe_total;
|
||||
|
||||
void vk_pipeline_prewarm_progress(VkPrewarmProgress *p) {
|
||||
p->phase = __atomic_load_n(&pw_phase, __ATOMIC_ACQUIRE);
|
||||
gfx_shader_prewarm_progress(&p->sh_done, &p->sh_total, &p->sh_translated);
|
||||
p->pipe_done = __atomic_load_n(&pw_pipe_done, __ATOMIC_RELAXED);
|
||||
p->pipe_total = __atomic_load_n(&pw_pipe_total, __ATOMIC_RELAXED);
|
||||
}
|
||||
|
||||
static void prewarm_run(void);
|
||||
|
||||
void vk_pipeline_prewarm(void) {
|
||||
prewarm_run();
|
||||
__atomic_store_n(&pw_phase, 3, __ATOMIC_RELEASE);
|
||||
}
|
||||
|
||||
static void prewarm_run(void) {
|
||||
const char *gpu = rt_config()->gpu;
|
||||
uint32_t pr_hist[16] = {0}, fs_hist[4] = {0}; // library state of the recorded pipelines (lb_start)
|
||||
if (!A.on || (gpu && !strcmp(gpu, "null"))) return;
|
||||
@@ -2609,6 +2628,7 @@ void vk_pipeline_prewarm(void) {
|
||||
int cpus = (int)plat_cpu_usable();
|
||||
const char *pe = cfg_get("NX2NAVI_PREWARM_XLATE"); // no fixed default: half the CPUs, at most 16, when unset
|
||||
int procs = pe && *pe ? atoi(pe) : cpus / 2 > 16 ? 16 : cpus / 2;
|
||||
__atomic_store_n(&pw_phase, 1, __ATOMIC_RELEASE);
|
||||
gfx_shader_prewarm((const uint8_t(*)[32])ids, u, sh, cpus, procs < 1 ? 1 : procs);
|
||||
double t1 = now_ms();
|
||||
|
||||
@@ -2656,14 +2676,24 @@ void vk_pipeline_prewarm(void) {
|
||||
plat_mutex_unlock(&P.lock);
|
||||
free(ids), free(sh);
|
||||
|
||||
__atomic_store_n(&pw_pipe_done, 0, __ATOMIC_RELAXED);
|
||||
__atomic_store_n(&pw_pipe_total, (int64_t)queued, __ATOMIC_RELAXED);
|
||||
__atomic_store_n(&pw_phase, 2, __ATOMIC_RELEASE);
|
||||
PlatThread *t[64];
|
||||
int nt = 0;
|
||||
for (int i = 0; i < cpus && i < 64 && queued; i++)
|
||||
if ((t[nt] = plat_thread_new(warm_builder, NULL, "pipe prewarm"))) nt++;
|
||||
if (!nt) warm_builder(NULL);
|
||||
double logged = now_ms();
|
||||
for (int i = 0; i < nt; i++)
|
||||
while (!plat_thread_join_ns(t[i], 2000000000u))
|
||||
rt_log("vk_pipeline: prewarm: %u of %u pipelines, %.0f ms", K.built + K.failed, queued, now_ms() - t1);
|
||||
while (!plat_thread_join_ns(t[i], 100000000u)) { // every 100 ms the numbers, every 2 s a log line
|
||||
uint32_t fin = __atomic_load_n(&K.built, __ATOMIC_RELAXED) + __atomic_load_n(&K.failed, __ATOMIC_RELAXED);
|
||||
__atomic_store_n(&pw_pipe_done, (int64_t)fin, __ATOMIC_RELAXED);
|
||||
if (now_ms() - logged < 2000) continue;
|
||||
logged = now_ms();
|
||||
rt_log("vk_pipeline: prewarm: %u of %u pipelines, %.0f ms", fin, queued, now_ms() - t1);
|
||||
}
|
||||
__atomic_store_n(&pw_pipe_done, (int64_t)queued, __ATOMIC_RELAXED);
|
||||
plat_mutex_lock(&P.lock);
|
||||
while (K.left) plat_cond_wait(&done_cv, &P.lock); // jobs a pool worker took
|
||||
rt_log("vk_pipeline: prewarm: %zu records in %s: %u pipelines built, %u failed, %u already built, %u dropped "
|
||||
|
||||
Reference in New Issue
Block a user